Independent researcher with a habit of reverse-engineering things that probably don’t want to be reversed. Currently obsessed with multilingual models, mechanistic interpretability, and AI safety.

GitHub · LinkedIn ·